406 



Dr. H. Herwig on the Expansion 



Therefore^ W 2 , W, R, and c once calculated for a series of 

 observations, the rest of the calculation is performed with the 

 two formulae 



^ = [W 2 (l+/30-W 1 (l +7 0] 



R 



and 



W,(l + |80-(W-W I )(l+70 



} 



dt 



dt ' 



It will now be seen from the following that the first term in 

 the right-hand brackets for 4>, on account of the small value of 

 H in the greatest number of my series of observations, is deci- 

 dedly the smaller of the two bracketed terms. An at all events 

 small error in the determination of / would, on this account, 

 have absolutely no influence ; hence this formula was especially 

 advantageous for the calculation of the experiments. 



§3. 



Before communicating the observations on vapours, I will first 

 cite two tests of accuracy to which I subjected the method. 

 First, I filled a tube on both sides with dry air ; constant values 

 of (f> must evidently then be obtained. The two following 

 Tables contain the results from two tubes. On their arrange- 

 ment nothing more need be added after what has been said. 



W = 307-6. 



Table I. 

 W=224-5. R=414-674. c=0'88. 



t. 



H. 



W v 



v. 



V. 



<p. 



P- 





millims. 











millims. 



307 



+4635 



89-59 



217-75 



1721 



557-90 



778 



42-5 



46-45 



89-5 



2177 



1719 



557-21 





52-3 



46-45 



89-5 



217-65 



171-7 



556-73 





65 



46-65 



89-31 



217-7 



171-3 



557-04 



865 







Table 11 



[. 





w 2 =c- 



509-6. 



W=224'5. R 



= 387-472. c= 



=0-88. 



16-3 



55-8 



81-24 



228-25 



16605 576-64 



731 



28-2 



56-1 



80-98 



228-4 



165-55 



57711 





42-4 



56-3 



80 8 



228-5 



1651 



577-05 





58-7 



56-4 



80-71 



228-4 



164-7 



576-17 



837 



The constancy of (j> is sufficiently attained ; it is seen, however, 

 within what limits small variations occur. 



